Big data en el análisis económico-financiero de la empresapropuestas empíricas en la predicción del fracaso
- NOGUERA VENERO, JOSÉ
- Manuel Ruiz Marín Director
- Mari Luz Maté Sánchez de Val Co-director
Defence university: Universidad Politécnica de Cartagena
Fecha de defensa: 17 March 2023
- Mariano Matilla García Chair
- Fernando Antonio López Hernández Secretary
- Mª Carmen Martínez Victoria Committee member
Type: Thesis
Abstract
If we imagine the market as a circuit in which different agents act among themselves to obtain an economic benefit, we can say that companies form part of this circuit together with other agents (such as households, the state, banks, investors, etc.). If any of these agents cease to exist, from an economic perspective, this affects those who remain in the circuit (deflation, unemployment, decrease in productive activity), and even other agents of the same type (loss of suppliers), even if they belong to other countries (decrease in international trade). In this sense, and because we are located in a capitalist system, we analyse the situation in which companies cease to exist because they are of fundamental importance in political and socio-economic spheres. The objective of this thesis is to determine the combination of necessary financial conditions that cause firms to fail. In addition, we examine whether these conditions vary in the face of unexpected and hard-to-predict external shocks such as the crisis resulting from the Covid pandemic situation. To this end, we develop a methodology based on symbolic analysis, from which we characterise the process of business failure based on the information we obtain from financial ratios. In particular, we overcome the limitations derived from the application of fuzzy set comparative qualitative analysis (fsQCA) techniques with respect to their subjective character by applying Big Data algorithms based on symbolic analysis, and the use of traditional statistical tools. In addition, we apply clustering tools in order to identify common financial characteristics in the different companies that we determined as failures in the previous process. This helps us to identify different processes of business failure. This proposed procedure is then applied to test whether the typology of business failure processes has changed before and after Covid. Subsequently, we develop an empirical analysis applying the previously described procedure to the primary sector, finding results that allow us to identify differences between different types of business failure processes. Finally, we test for the existence of territorial clustering processes for the different types of business failure processes and whether there exist differences in the spatial distribution when we compare pre- and post-Covid periods. This thesis contributes to the business failure literature with a procedure to identify different types of business failure processes based on the financial information of the company